US orders diplomats to raise concerns over alleged Chinese AI theft

Washington: The US State Department has directed its diplomatic missions worldwide to press foreign governments on what it describes as broad efforts by Chinese companies, including artificial intelligence startup DeepSeek, to obtain intellectual property from US AI laboratories, according to a diplomatic cable seen by Reuters.

The cable, dated Friday and sent to diplomatic and consular posts globally, instructs US personnel to engage their counterparts on concerns about adversaries extracting and distilling American AI models. The document also says, A separate demarche request and message has been sent to Beijing for raising with China.

Distillation refers to the practice of training smaller AI systems using the outputs of larger, more costly models in an effort to reduce the expense of building advanced AI tools.

The move follows similar allegations made by the White House earlier this week. The State Department did not immediately respond to a request for comment, according to Reuters.

DeepSeek and other firms named

OpenAI has warned US lawmakers that DeepSeek was targeting the ChatGPT developer and other leading American AI companies in an effort to reproduce their models and use them in its own training.

The State Department cable said its purpose was to warn of the risks linked to using AI models distilled from proprietary US systems and to prepare the ground for possible further outreach by the US government. It also named Chinese AI firms Moonshot AI and MiniMax. Neither company immediately responded to a request for comment.

AI models developed from surreptitious, unauthorised distillation campaigns enable foreign actors to release products that appear to perform comparably on select benchmarks at a fraction of the cost but do not replicate the full performance of the original system.

The cable further said such campaigns also deliberately strip security protocols from the resulting models and undo mechanisms that ensure those AI models are ideologically neutral and truth-seeking.

China rejects the allegations

The Chinese Embassy in Washington rejected the accusations on Friday and repeated Beijing’s position that the claims have no basis.

“The allegations that Chinese entities are stealing American AI intellectual property are groundless and are deliberate attacks on China’s development and progress in the AI industry”, the embassy said in a statement to Reuters.

DeepSeek did not immediately respond to a request for comment. The company has previously said its V3 model relied on naturally occurring data gathered through web crawling and that it had not intentionally used synthetic data generated by OpenAI.

On Friday, DeepSeek released a preview of a new model, V4, adapted for Huawei chip technology, highlighting China’s increasing self-reliance in the sector. The company’s lower-cost AI model had drawn global attention last year.

Many Western governments, along with some in Asia, have barred institutions and officials from using DeepSeek over data privacy concerns. Even so, DeepSeek’s models have remained among the most widely used on international platforms hosting open-source models.

The White House allegations and the State Department cable come only weeks before US President Donald Trump is due to visit Chinese President Xi Jinping in Beijing. Reuters said the developments could add strain to a prolonged technology rivalry between the two countries after tensions had eased under a detente reached last October.